This invention relates to laser doppler velocimeters and in particular to a laser doppler velocimeter for measuring wind velocity such as to ascertain true air speed or to measure gradients such as wind shear and having reduced cost and weight and increased eye safety.
Laser doppler velocimetry (LDV) in gases at long ranges has been a subject of investigation for nearly 20 years. For instance U.S. Pat. No. 3,915,572 issued Oct. 28, 1975, discloses a laser Doppler velocimeter. The basic LDV process compares a reflected beam of light to a reference beam to determine the Doppler shift in the reflected beam. This Doppler shift is then used to measure air speed or gradients such as wind shear. Early velocity measurements were performed in large wind tunnels at distances of a few meters. Subsequent investigations included the measurement of meteorological parameters including wind velocity and turbulence, in part because of their importance to aircraft flight. By the early 1970's, laser doppler velocimeters were operating at ranges of hundreds of meters and, by the early 1980's, measurements were being made at kilometer ranges. These systems were large and weighed hundreds of pounds.
A typical long-range LDV for measuring wind shear includes a source of polarized coherent radiation, such as a CO.sub.2, YAG, or argon laser (preferably lasing in the fundamental transverse mode and in a single longitudinal mode), for projecting a first coherent beam of light into a beam shaper. The beam shaper expands and collimates the beam after which the beam enters a telescope. The telescope projects the beam in nearly collimated form.
The telescope focuses the beam to strike airborne particulates at a point of interest, resulting in a scattered beam. A portion of the scattered beam is reflected by aerosols and received either by the telescope or a separate receiver telescope. The scattered beam is then mixed with a separate reference beam of light in an optical mixer.
The optical mixer is typically a photodetector with bandwidth sufficient to detect the Doppler frequency as described below. The reference beam is selected to have a well defined optical frequency. If the particulates are in motion relative to and along the axis of the telescope, then the scattered beam has a frequency that is shifted by an amount equal to the Doppler frequency.
The mixing process produces an electrical current containing a component whose frequency is the mathematical difference between the frequency of the reference beam and the frequency of the Doppler shifted scattered beam. The difference frequency can therefore be measured by electrical means such as by use of an electrical spectrum analyzer or by use of a frequency counter. Because the Doppler frequency is proportional to the relative velocity component of the telescope and the particulates along the telescope axis and because the proportionality constant is a precise mathematical constant determined entirely by the wavelength of the emitted optical beam, the relative velocity component of the particulates along the telescope axis can be determined once the Doppler frequency is measured.
The reference beam is selected to have a well-defined optical frequency. In most cases, for convenience, the reference beam is a derivative or replica of the original transmitted beam. If the reference frequency is equal to the frequency of the transmitted beam, then the mixing process is referred to as homodyne detection and the resulting electrical frequency is equal to the Doppler frequency. If the reference frequency is shifted by a known constant amount, then the mixing process is referred to as heterodyne detection and the electrical current has a frequency equal to the Doppler frequency plus a constant offset frequency. In either case, the Doppler frequency can be measured.
The mixing approach described above, which takes either the heterodyne or homodyne forms, is not only a convenient means of extracting the Doppler signal from the scattered optical wave, but is also the most sensitive means of detecting a very weak optical signal. The electrical current produced by the mixer contains two components that are proportional to the optical power in the reference beam and the scattered beam, respectively, and a third component containing the Doppler signal that is proportional to the geometric mean of the powers in the two optical beams. Other sources of electrical current, such as leakage currents in the detector and Johnson and shot noise currents, may also be present.
Typically, all of the electrical currents present in the mixer current can be viewed as steady in comparison to the Doppler signal current. The detection of the Doppler signal current therefore involves measurement of a time varying current among steady background currents. Because the scattered beam power is very weak and because the telescope collects only a small fraction of what is scattered, this time varying optical signal is small in comparison to the reference beam. The mixing process helps improve detection sensitivity of the small optical signal.
The Doppler electrical current experiences "heterodyne gain" because the mixer forms the mathematical product of the scattered optical wave and the reference optical wave in generating the current. By using a reference beam of sufficient power, the only significant detected electrical currents are the steady current corresponding to the directly detected reference beam and the time varying current containing the Doppler signal.
The detection sensitivity is then determined by the noise characteristics of the reference beam. Therefore, in general, it is of utmost importance that the reference beam be as stable as possible. In practice, very stable laser sources can produce reference beams whose power fluctuations are only several decibels above the shot noise floor. Use of a shot noise limited reference beam can be shown to produce a detection sensitivity in the Doppler mixer that is shot noise limited. This kind of detection arrangement is sometimes referred to as quantum limited detection. With modern laser sources it is possible to come within several decibels of this sensitivity limit using the mixing approach described.
Despite the sensitivity limits involved in optical mixing, the return signals produced by optical scattering from aerosols are so small that special consideration must be given to the optical power transmitted from the source laser. Power considerations have resulted in use of sources such as YAG or CO.sub.2 lasers, running Q-switched or continuously, in laser Doppler velocimeters. However, the signal-to-noise level at the optical mixer is influenced by two factors: detection bandwidth and signal level. Larger signal levels are produced when more optical power is transmitted and therefore signal-to-noise is improved by using larger amounts of transmitted optical power. By using longer transmitted pulses of optical power, the frequency bandwidth of the detected signal may be reduced so that noise filtering can be made more effective, and, in turn, the signal-to-noise ratio improved.
Longer pulses may only be used if the reference signal and the scattered Doppler signal are phase coherent over the time duration of the transmitted pulse, which is the case for the laser system of the present invention. The signal-to-noise level at the optical mixer is therefore proportional to the product of transmitted optical power and transmitted pulse duration, or equivalently, optical energy per pulse.
The source of the reference beam is often a major problem in this type of system. Originally, the reference beam is mutually coherent with the output beam. However, over the transit time of the launched beam, the two beams can become decorrelated, which can result in significant measurement error. To overcome this problem in some systems long coherence length lasers are used so that one simply mixes a portion of the laser light split from the source laser with the return wave to attain interference. Alternately, a second laser can be used to generate the reference beam provided that it can be properly phase-locked to the source laser. The laser system described here has sufficient coherence to enable use of the former approach even when the focal distance is as large as several miles.
Another major limitation in the use and application of laser systems to the measurement of wind shear or air speed has been the lack of an eye-safe source of radiation with sufficient energy. The Army medical standard recognizes a relatively high maximum permissible exposure in joules/cm.sup.2 at wavelengths in approximately the 1.51-1.56 micron range. YAG and CO.sub.2 lasers do not operate in this wavelength range, and thus the power levels and energy per pulse levels necessary for eye-safe operation of these lasers are low.
Lasers based on erbium-doped glass emit radiation in this eye-safe band of wavelengths. Therefore, military laser target range finders and laser target designators are now using flash-pumped erbium laser sources. These systems, like most erbium lasers to date, are based on an erbium-doped glass rod lasing medium that is excited by using a flash pump. Although they operate at a desirable wavelength, erbium lasers are flash-pumped and include discrete optical elements that have alignment, thermal drift, and vibration problems.
Recent developments in the fiber optics field have resulted in lightweight, low cost, highly efficient, erbium-doped glass fiber. These systems can be efficiently pumped by using recently developed laser diodes whose emission wavelengths have been optimized for excitation of erbium-doped glass. The overall laser diode pumped erbium fiber system also eliminates the cooling requirements associated with previous flash-pumped systems. The development of erbium doped fibers has lowered the cost and much of the bulk and optics associated with the use of rods. In addition, fibers, like rods, may be used as amplifiers, energy storage devices, or as the lasing medium in a fiber laser. However, use of optical fiber technology has limitations. Specifically, the fiber has inherent power limitations and is subject to electric field-induced non-linearities at higher optical powers. Using erbium-doped fiber as the laser in a LDV system has not been practical because to obtain the power in the fiber, a large amount of energy must be stored and then the laser Q-switched. The Q-switching leads to uncontrolled dumping of the energy, which can introduce electrical field-induced non-linearities at higher optical powers or could ultimately damage the fiber.